The dilemma in treating men with castration resistant prostate cancer (CRPC) lies not only in the heterogeneity of the disease but also the spectrum of patients that have the disease. Considerable energy has been dedicated to understanding tumor heterogeneity and developing prognostic models of clinical outcomes in men with CRPC who are chemotherapy naive. The identification of prognostic factors of clinical outcomes in men with CRPC who failed frontline chemotherapy has not, however, been investigated. Such identification is increasingly important due to the large number of patients who not only fail frontline chemotherapy but have excessive toxicity due to docetaxel. We anticipate that the same patient may need different models (prognostic calculators) at different stages of their care pathway as more prognostic information on that person's condition accumulates. The specific aims in this proposal are: 1) to develop a prognostic model that will predict overall survival (OS) in men with CRPC who failed first line chemotherapy. The model will be validated for predictive accuracy using an independent dataset. 2) To develop a prognostic model that will predict progression-free survival (PFS) in CRPC men who failed first line chemotherapy. The model will be validated for predictive accuracy using an independent dataset. 3a) To determine if e 30% decline in prostate specific antigen (PSA) at 3-months following treatment with cabazitaxel, the only FDA approved drug for treating men who failed frontline chemotherapy, is a valid surrogate marker of OS. 3b) to develop and validate a prognostic model that will predict post-therapy decline in PSA (e30% decline from baseline at 3-months). 4) To test for the dependence between time to progression and OS using the TROPIC trial. Innovation: This study has a high degree of innovation because of its strong potential to positively impact the design and conduct of future trials in PC. In particular, this study will be the first to identify and validate models of clinical outcomes and will incorporate data from the two largest phase III trials of men with advanced CRPC who failed frontline chemotherapy. These models are completely absent for the growing group of men who have received frontline chemotherapy and are considering secondary chemotherapy. The development of these models will facilitate discussions with CRPC patients, as well as help integrate these models into the design, conduct and analysis of new clinical trials in PC and patient care. PUBLIC HEALTH RELEVANCE: Prognostic models are not available for the burgeoning group of men who have received one regimen of chemotherapy, and are considering secondary chemotherapy. If these results are validated, our models can be used prospectively in randomized phase II &phase III trials to ensure that the treatment groups being evaluated are comparable. Further, they can assist in identifying subgroups of men with CRPC with particularly good or poor prognoses for whom therapy can be subsequently tailored. These models will serve as prediction tools that can be easily and rapidly implemented in clinical practice.